Joint Carrier Phase Estimation and Symbol Decoding of Trellis Codes

We use the Expectation-Maximization algorithm to deduce an iterative method for carrier phase estimation and symbol decoding. The method applies to all trellis-based codes. Carrier phase offset and noise variance estimation is based on the maximum likelihood criterion. The Markovian properties of the encoder state sequence enable us to calculate the required likelihoods using a forward-backward algorithm. Given the phase estimate, the calculated likelihood functions can easily give optimum decisions on information symbols which minimize the symbol error probability. Simulation results which show rapid convergence are given.